| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 22.1 | It has been considered | DACT19::PDORNAN | Patrick Dornan, NWSS 8-339-7169 | Thu Dec 21 1989 12:58 | 6 |
| At FDDI PID training held for NWSS on 12/18, this was discussed.
The first FDDI to Ethernet Bridge will have the potential for this
type of problem, but it will be fixed with a firmware update soon
after.
Patrick
|
| 22.2 | | KONING::KONING | NI1D @FN42eq | Thu Dec 21 1989 13:46 | 14 |
| Huh?
Clearly you can't crank 100 megabits of stuff through a 10 megabit
pipe, so the scenario in .0 can obviously occur. But I suspect the
bigger problem with remote bridges such as those made by Vitalink is a
different one: under high loads, they often stop processing the Bridge
Hello messages, which causes the entire bridge topology to get
confused.
That sort of problem IS handled in DEC bridges in general and the
10/100 bridge in particular: Bridge Hello messages are always
processed, independent of possible congestion on data packets.
paul
|
| 22.3 | Problem still there | LASHAM::PRUDEN_G | | Fri Dec 22 1989 05:43 | 22 |
| Yes disguarding BPDUs was one of the problems with Vitalink but it has
been fixed (or so Vitalink say) in their latest release of software.
However, as soon as the forwarding rate of one of these bridges
is reached the problem described in .0 still occurs. FDDI bridges
will have the same problem since, as .2 points says, there is a
mismatch of speeds.
In the info that we have been putting out no one has even mentioned
the problem. I beleive that the problem can be solved by making sure
that the traffic on the Ethernet is carefully monitored and when
it becomes high push the FDDI ring furthur out i.e. reduce the size
of Ethernets by putting more workstations on FDDI or split them into
smaller Ethernets.
This however assumes that good management is being adopted and
a customer knows what traffic levels are on the Ethernets. I know
several customers who have very, very large Ethernets who will be
going to FDDI. These Ethernets today are badly managed and the bridge
problem pointed out in .0 is being seen. They see FDDI as the answer
but probably won't improve their management. It seems to me that
we should be pointing these sorts of problems out to them so they
don't end up with similar problems on FDDI.
Gary
|
| 22.4 | This is not intended to start us down a rathole | STAR::SALKEWICZ | It missed... therefore, I am | Tue Jan 02 1990 12:04 | 23 |
| re .3 FDDI/Ethernet Bridge throughput considerations 3 of 3
Hi Gary,
While I do agree with everything you are saying, and Paul
too for that matter,.. I feel complelled to add this little quip...
> It seems to me that
> we should be pointing these sorts of problems out to them so they
> don't end up with similar problems on FDDI.
It seems to me that we have been telling our customers to do good
management all along,.. and some chose to disregard our warnings/advice,..
to the point where they allow their networks to degrade into useless
piles of hardware. To tell them about this potential future problem
will oprobably not cause them to act, since they do not act on the
present problems caused by bad management that exist today. It is
unfortunate that people do not take network management more seriously,...
and I think this point touches on that larger problem.
You can lead a horse to water ....
/Bill
|